In circuits in which digital signal processing is performed, data is typically handled in the form of binary words, which are transferred on data busses. In certain types of circuits, the digital signal may be a digital representation of an ideally continuous quantity, such as an analog signal. When digital signal processing is employed, the continuous nature of the signal may be lost, in that digital words may only be increased or decreased in step increments. In general, the size of the smallest step increment is inversely proportional to the available signal resolution.
One way to increase resolution in digital signal processing systems is to increase the number of bits in the digital signals processed. For example, a six bit system has the capability for more resolution than a four bit system, but does not have the capability for as much resolution as an eight bit system. One drawback of increasing the number of bits in a digital signal to increase the resolution capability is that circuitry for the manipulation of higher bit signals is more complex and costly than similar circuitry for the manipulation of lower bit signals.
In certain systems, absolute high resolution is not necessary to increase system performance. Instead, a means which would alleviate large step changes is suitable to provide the system with apparent high resolution, without incurring the cost of high resolution digital circuitry.